Hi All,
Just wondering what I am doing wrong here, as it seems like backup of promox might be trying to backup a NFS mount, but documentation/forums imply otherwise.
I have a LXC, that has a NFS mount to a NAS inside it, this works great. However I also tried to have 'daily snapshot' backups on all the machines, just so I have a rollback point. This works create with everything else, but falls over on the LXC.
If I unmount the NFS mount (inside the LXC) this works perfectly, but when mounted, the backup task locks the machine for 6+ hours, and never really completes. I can only assume its trying to backup the entire NFS NAS it's connected to, which is quite big.
I've noticed some people create the NFS mount in proxmox itself, and then use a bind point, that might work, but at the moment I have all the machines in a 3 node cluster with CEPH, so don't want to have to create a NFS mount on every host (and then risk not doing it correct on anyone).
Is there a way to tell the backup scheduler, in a cluster, to say avoid the /mnt folder?, or ideally just to backup the filesystem/disk image itself, and not any remote mounts?. Or is there anyone who understands why having NFS mounted in a LXC would cause the backup tasks to fall over completely?
Just wondering what I am doing wrong here, as it seems like backup of promox might be trying to backup a NFS mount, but documentation/forums imply otherwise.
I have a LXC, that has a NFS mount to a NAS inside it, this works great. However I also tried to have 'daily snapshot' backups on all the machines, just so I have a rollback point. This works create with everything else, but falls over on the LXC.
If I unmount the NFS mount (inside the LXC) this works perfectly, but when mounted, the backup task locks the machine for 6+ hours, and never really completes. I can only assume its trying to backup the entire NFS NAS it's connected to, which is quite big.
I've noticed some people create the NFS mount in proxmox itself, and then use a bind point, that might work, but at the moment I have all the machines in a 3 node cluster with CEPH, so don't want to have to create a NFS mount on every host (and then risk not doing it correct on anyone).
Is there a way to tell the backup scheduler, in a cluster, to say avoid the /mnt folder?, or ideally just to backup the filesystem/disk image itself, and not any remote mounts?. Or is there anyone who understands why having NFS mounted in a LXC would cause the backup tasks to fall over completely?